Network analysis of mineralogical systems

Shaunna M. Morrison, Chao Liu, Ahmed Eleish, Anirudh Prabhu, Congrui Li, Jolyon Ralph, Robert T. Downs, Joshua J. Golden, Peter Fox, Daniel R. Hummer, Michael Meyer, Robert M. HazenAmerican 2017 · DOI

Abstract

A fundamental goal of mineralogy and petrology is the deep understanding of mineral phase relationships and the consequent spatial and temporal patterns of mineral coexistence in rocks, ore bodies, sediments, meteorites, and other natural polycrystalline materials. The multi-dimensional chemical complexity of such mineral assemblages has traditionally led to experimental and theoretical consideration of 2-, 3-, or n-component systems that represent simplified approximations of natural systems. Network analysis provides a dynamic, quantitative, and predictive visualization framework for employing "big data" to explore complex and otherwise hidden higher-dimensional patterns of diversity and distribution in such mineral systems. We introduce and explore applications of mineral network analysis, in which mineral species are represented by nodes, while coexistence of minerals is indicated by lines between nodes. This approach provides a dynamic visualization platform for higher-dimensional analysis of phase relationships, because topologies of equilibrium phase assemblages and pathways of mineral reaction series are embedded within the networks. Mineral networks also facilitate quantitative comparison of lithologies from different planets and moons, the analysis of coexistence patterns simultaneously among hundreds of mineral species and their localities, the exploration of varied paragenetic modes of mineral groups, and investigation of changing patterns of mineral occurrence through deep time. Mineral network analysis, furthermore, represents an effective visual approach to teaching and learning in mineralogy and petrology.

Introduction

Mineralogy and petrology deal with immense chemical complexity. Traditionally, researchers have understood mineral relationships through simplified 2- or 3-component phase diagrams, such as ternary plots. While foundational, these methods struggle to represent natural systems where dozens of minerals might coexist across thousands of unique localities. As mineralogical databases grow into the realm of "big data," there is a pressing need for a quantitative framework that can handle high-dimensional relationships. This paper introduces network analysis—tools more commonly found in sociology or protein-protein interaction studies—to explore how minerals are distributed in space and time.

Methodology

The authors represent mineral systems as mathematical graphs where nodes (vertices) represent individual mineral species and links (edges) represent the coexistence of two minerals at the same location or in the same rock sample. Key techniques include:

  • Fruchterman-Reingold Force-Directed Layouts: An algorithm that treats edges like springs and nodes like repelling magnets, naturally causing highly connected minerals to form "cliques" or clusters based on their shared environments.
  • Multi-Dimensional Scaling (MDS): A technique used to project high-dimensional coexistence data into 2D or 3D space while preserving the relative "distances" (inverse of coexistence frequency) between minerals.
A force-directed network graph of common rock-forming minerals. The clustering (colored nodes) emerges naturally from coexistence data, reflecting different paragenetic environments.

The study utilizes high-quality datasets, including classic petrological records of igneous rocks and the massive, crowd-sourced database Mindat.org, which provides coexistence data for thousands of global localities.

Key Insights and Results

Network analysis reveals that the structure of mineral systems mirrors underlying thermodynamics. Traditional ternary phase diagrams and Bowen’s reaction series emerge naturally from the network topology. For example, MDS plots can recover the sequential pathways of mineral formation.

2D Multi-Dimensional Scaling (MDS) plot of rock-forming minerals. The arrows highlight pathways that correspond to established mineral reaction series, demonstrating how high-dimensional data recovers traditional petrological insights.

Furthermore, the networks provide a visual record of Earth's evolution. By comparing copper networks from the Archean Eon to the Cenozoic Era, the authors visualized the impact of the Great Oxidation Event, showing a massive explosion of oxidized species (sulfates, carbonates) in the more recent network. Quantitatively, igneous rock-forming minerals form a dense, compact network (Density = 0.64), while rarer groups like chromium minerals form sparser networks (Density = 0.05), reflecting more diverse and isolated formation environments.

Conclusion and Future Directions

Mineral network analysis provides a powerful new lens for geology, offering predictive power to guide the discovery of new mineral species or ore deposits. Future work will extend these tools to planetary science, comparing the mineralogical evolution of Earth with Mars, the Moon, and other planetary bodies. While the quality of the network depends on the accuracy of the input data—specifically distinguishing true equilibrium assemblages from accidental proximity—this framework represents a significant leap forward in managing the complexity of the mineral kingdom.

Summary updated

Abstract

A fundamental goal of mineralogy and petrology is the deep understanding of mineral phase relationships and the consequent spatial and temporal patterns of mineral coexistence in rocks, ore bodies, sediments, meteorites, and other natural polycrystalline materials. The multi-dimensional chemical complexity of such mineral assemblages has traditionally led to experimental and theoretical consideration of 2-, 3-, or n-component systems that represent simplified approximations of natural systems. Network analysis provides a dynamic, quantitative, and predictive visualization framework for employing "big data" to explore complex and otherwise hidden higher-dimensional patterns of diversity and distribution in such mineral systems. We introduce and explore applications of mineral network analysis, in which mineral species are represented by nodes, while coexistence of minerals is indicated by lines between nodes. This approach provides a dynamic visualization platform for higher-dimensional analysis of phase relationships, because topologies of equilibrium phase assemblages and pathways of mineral reaction series are embedded within the networks. Mineral networks also facilitate quantitative comparison of lithologies from different planets and moons, the analysis of coexistence patterns simultaneously among hundreds of mineral species and their localities, the exploration of varied paragenetic modes of mineral groups, and investigation of changing patterns of mineral occurrence through deep time. Mineral network analysis, furthermore, represents an effective visual approach to teaching and learning in mineralogy and petrology.
BibTeX
@inproceedings{morrison2017network,
  author = {Shaunna M. Morrison and Chao Liu and Ahmed Eleish and Anirudh Prabhu and Congrui Li and Jolyon Ralph and Robert T. Downs and Joshua J. Golden and Peter Fox and Daniel R. Hummer and Michael Meyer and Robert M. Hazen},
  title = {{Network analysis of mineralogical systems}},
  booktitle = {American},
  year = {2017},
  doi = {10.2138/am-2017-6104ccbyncnd},
}
Figure